Food Photography Prompts for Restaurant and Cookbook Stills

Food photography prompts name the dish, plating, camera height, light, and table surface so Midjourney v7, FLUX, Leonardo, SDXL, or GPT Image draw restaurant menus and cookbook stills you can brief a kitchen or a designer. You want paste-ready language for overhead bowls, 45-degree burgers, dark-slate mains, and steam-forward close-ups. You leave with a food still stack, restaurant and cookbook kits, dialect notes as of mid-2026, and a photo-first path when a plate photo exists. This page stays on plated food. This guide supports the food photography use case on PromptMake. Soft path when a reference still holds the plate: PromptMake /image, then confirm angle and garnish before you generate.

Who needs food photography prompts

You type food photography prompts because a model with no plate instruction picks a default: eye-level snack crop, mixed garnish, mystery table. That default can serve a casual snapshot. It fails a tasting-menu hero and a cookbook spread that must match a chapter. Chefs and restaurant marketers need a locked overhead for bowls and a 45-degree camera for stacked burgers. Cookbook teams need a consistent surface and key so pasta on page 40 matches dessert on page 88. Name the dish, plating style, camera height, light source, and table surface in the prompt. Vague "delicious food" returns stock tropes and plastic shine.

Text generation invents a plate from nouns. Image-to-prompt tools that read a still own the "keep this angle and surface, rewrite the dish" job. This article covers writing food language: a five-part stack, kits for restaurant menus and cookbook spreads, Midjourney v7 and FLUX paste dialects, and photo-first Recreate or Create Variation when you upload a plate. Soft product path for the photo lane: https://promptmake.net/image. For kits tuned to food vocabulary, the food photography use case sits at https://promptmake.net/use-case/food-photography. Camera height belongs in the food stack here; a sister post on this blog owns general camera kits. White-sweep packshot work belongs in the product-photo article.

Treat generator stills as menu tests and recipe concepts. Clear location and likeness rights before you publish a chef or a dining room.

Strong fit for:

• Restaurant and hotel teams who need menu stills with a locked plate and surface
• Cookbook and editorial teams who need chapter-consistent light and props
• Food marketers who need campaign plates without a full studio day
• Teams who port the same plate line into Midjourney v7, FLUX, SDXL, or GPT Image

The food still stack: dish, angle, light, surface

Treat food photography prompts as a short plate order. Models fill a missing fact with stock tropes: overhead on a burger that needed height, harsh top light on chocolate. Write five facts in a stable order. You name the dish with cuisine and one hero texture: seared salmon skin or glossy ramen broth. You name plating as vessel and garnish: wide ceramic bowl, herb finish. You name camera height as the viewpoint: overhead flat lay, 45-degree three-quarter, eye-level hero, or macro on a cut. You name light as source and quality: window side light or warm tungsten from camera left. You name surface as table and linen: dark walnut, honed marble, slate. Write the stack once as labeled facts. Format twice: compact phrases plus trailing parameters for Midjourney v7 (and V8 Alpha where you have access), flowing sentences for FLUX family builds. SDXL and Leonardo take denser food tokens. GPT Image prefers a clean paragraph you can revise in chat. Leave ecommerce packshot language out so a white sweep does not fight a restaurant table.

Dish, plating, and cuisine lock

Open with the food, not a mood adjective. "Seared duck breast, cherry gastrique" beats "gourmet dinner." Add cuisine as a lock so the model does not drift: rustic Italian, kaiseki, bistro French, regional Mexican. One cuisine per prompt. Plating is the vessel plus garnish count. A tasting-menu plate wants negative space, a sauce swipe, and two garnish nouns. A trattoria bowl wants volume, grated cheese, and torn herb. Name the ceramic: wide matte bowl, coupe plate, slate board, cast-iron skillet. "Artisan plate" is weaker than "matte speckled ceramic, 28cm." Steam, condensation, and a cut face (lasagna layers, cake crumb) are freshness cues you write as short clauses. Skip piles of taste words. The generator draws light and texture; it does not taste. For Midjourney v7, keep the dish clause first: "restaurant food photography, pappardelle ragu, wide ceramic bowl." For FLUX, put the same facts in a sentence: Photograph of pappardelle with ragu in a wide ceramic bowl, torn basil, grated parmesan.

Camera height, light, and surface

Camera height is the food shot. Overhead (nadir, camera above the table) fits bowls, pasta, mezze, and spreads where the layout is the story. A 45-degree three-quarter fits burgers, stacked desserts, roasted birds, and cocktails with height. Eye-level hero fits a slice, a pour, a fork lifting pasta. Macro holds a crumb, a glaze drip, a sesame plane. Pick one height. Light needs a source and a side: north window from camera left, warm tungsten key, low-key single lamp for chocolate and steak, bright window for salads and crudo. Overhead ceiling light flattens sauce and kills shadow on a burger bun. Surface sells the venue. Dark walnut and slate read as restaurant night. Pale oak and linen read as cookbook day. Honed marble sits between. Name linen or napkin once. Conflicting pairs waste a generate: "overhead flat lay, 45-degree burger hero, creamy 85mm bokeh" asks for three cameras. Pick one kit per run. Midjourney users add --style raw for photographic food. FLUX users spell the same optics in prose. Pair canvas after the camera line: --ar 4:5 for menu and cookbook pages, --ar 1:1 for many overheads, --ar 16:9 for a site hero.

Restaurant menus and cookbook spreads

Restaurant kits and cookbook kits share a habit: you lock surface and key before you swap the dish. Menu stills fail when the table changes from walnut to marble between the starter and the main. Cookbook spreads fail when page 12 is high-key linen and page 40 is a black void. Write a paste template per kit and reuse it across plates so the venue stays stable while the food changes. Midjourney v7 wants compact phrases and trailing --ar plus --style raw for photos. FLUX.1.x and Flux 2 family builds (name the variant your host shows) want a full sentence that embeds dish, height, light, and table. Keep --ar as canvas shape. "45-degree" and "window side light" are shot facts. The subsections below give paste-ready restaurant and cookbook cores you can drop a dish into. Hands and waiter motion belong in a lifestyle clause you add on purpose; leave them out of a clean menu plate. Soft path if you shot the pass on a phone: upload to PromptMake /image, Recreate Exactly to recover angle and surface, then tighten garnish by hand.

Restaurant and hotel menu stills

Restaurant hero (Midjourney v7): restaurant food photography, tasting-menu plate, seared duck breast, cherry gastrique swipe, wide matte ceramic, micro herbs, 45-degree camera, window side light from camera left, dark walnut table, steam rising --ar 4:5 --style raw --v 7

Restaurant hero (FLUX prose): Photograph of a tasting-menu plate, seared duck breast with cherry gastrique, 45-degree camera at chest height, soft window light from camera left, dark walnut table, matte ceramic, visible steam, commercial food photography.

Bowl overhead for menus: "overhead restaurant food photography, tonkotsu ramen, nori, chashu, ajitama, scallion, glossy broth, dark ceramic bowl, nadir camera, single warm key from camera left, dark wood --ar 1:1 --style raw --v 7." Hotel room-service and banquet menus want the same lock: one surface, one Kelvin, one height for the whole PDF. Fine-dining plates use negative space and a sauce gesture. Bistro plates can show more of the table and a linen fold. Keep extra cutlery out unless the brief is a place setting. Add "Michelin-style plating" when you want sparse composition; add "trattoria portion" when you want volume. Change the protein and sauce on reuse so a menu set shares one table. Leave cookbook crumb scatter and recipe cards out of this kit so a chapter prop does not land on a menu still.

Cookbook chapter and editorial spreads

Cookbook overhead (Midjourney v7): cookbook food photography, overhead flat lay, pappardelle with ragu, basil, parmesan, linen napkin, ceramic bowl, warm side light, rustic oak table --ar 1:1 --style raw --v 7

Cookbook process (FLUX): Cookbook photograph of a lasagna cut face, eye-level camera, visible layers, rustic oak table, linen napkin repeating from the pasta chapter, warm window light from camera left, editorial food photography.

A chapter kit is a palette: oak, linen, speckled ceramic, 4800K to 5200K window, overhead for bowls, 45-degree for stacked bakes. Write those nouns once and paste them under each recipe in the chapter. Process stills (mise, dough fold, cut face) use the same table so the how-to pages match the hero. Editorial spreads can add a prop story: wine glass stem and herb sprig. Cap props at two so the dish stays the subject. Vertical --ar 4:5 fits a print page. Square --ar 1:1 fits a recipe card. Keep restaurant tasting-menu negative space out of a family-cookbook brief if the art director asked for generous portions. Save the winning surface string beside the PNG (ch3_oak_linen_v2.png) so the next recipe does not reopen a marble table and break the chapter.

Overhead, 45-degree, and hero close-up kits

Three food heights cover most restaurant and cookbook work. Overhead is a map of the plate. Forty-five degrees is a portrait of height. Hero close-up is a texture crop. Mixing all three in one line produces a compromise tilt and a confused focus plane. Write altitude of the camera in plain words: camera above the table, camera at 45 degrees, camera at plate height, camera centimeters from the glaze. Name the focus noun (broth surface, bun sesame, cake crumb) so the model knows what must be sharp. Midjourney v7 responds to "overhead food photography" or "45-degree food photography" plus --style raw. FLUX responds to a sentence that includes camera height and table. Macro food uses the same parallel-plane habit as product macro: camera square to the cut, 90mm to 105mm class, thin focus. Keep 85mm portrait language and drone language out of food kits. The paste cores below are height-first. Drop your dish into the first clause and leave the height line alone across a set.

Overhead kit (Midjourney v7): overhead food photography, mezze spread, hummus, olives, pita, olive oil pools, nadir camera, warm side light, stone table --ar 1:1 --style raw --v 7

Forty-five degree kit (FLUX): Food photograph at 45 degrees of a smash burger, bun height visible, melted cheese drip, sesame plane sharp, window light from camera left, dark slate, restaurant food photography.

Hero close-up: "macro food photography, chocolate fondant cut, molten center, powdered sugar, camera parallel to the cut, dark moody background, low-key key from camera left --ar 4:5 --style raw --v 7." Burgers, layer cakes, and pours want 45-degree or eye-level. Soups, grain bowls, and cookie sheets want overhead. A steak with a sauce lake can go either way: overhead for the lake shape, 45-degree for the crust. Pick based on what the diner should notice first. SDXL users can add a short negative against plastic food, wax, and extra utensils. GPT Image lets you follow up with "move to 45 degrees, keep the walnut table."

Midjourney, FLUX, and GPT Image dialects

Porting a winner means translating the plate line. Midjourney v7 rewards a compact stack: dish, plating, height, light, surface, then --ar, --style raw for photos, --v 7. Style references (--sref) lock look inside the venue you chose; you write "45-degree" or "dark walnut" yourself. FLUX.1.x and Flux 2 family builds take natural-language prompts. Put the shot in the sentence: photographed at 45 degrees, window light from camera left, dark walnut table. Size sits in the host UI or API. SDXL wants food tokens in the positive prompt and a short negative when wax and plastic keep appearing. Leonardo remains a practical free-tier target; use the same height and surface nouns. DALL·E / GPT Image inside ChatGPT stay conversational: "make this a 45-degree restaurant plate on walnut" as a follow-up works because the thread holds the dish. Ideogram v3 helps when a recipe title must sit in frame; name the plate so type does not float on a mystery crop.

Language models can draft the stack before you spend a generate. As of mid-2026, GPT-5.6 Sol (API gpt-5.6) handles hard brief writing; ChatGPT defaults to GPT-5.5 Instant for fast chat. Claude Fable 5 and Claude Opus 5 tidy garnish lists. Ask for one prompt plus a separate line that states dish, height, light, and surface.

Step-by-step: lock the plate before you generate

A fixed order beats lucky garnish. Decide the job (restaurant menu still or cookbook chapter), write the five-part food stack, format for the model, generate once, then change one layer (garnish, grade, or dish) without touching height or table unless the brief changed. Two to five rounds with a locked surface beat ten rounds that hop from walnut to marble. If you hold a photo or a prior AI still with the correct plate, run image-to-prompt so the first draft inherits height and linen, then enforce the matching Kelvin on paste. PromptMake /image formats for Midjourney, FLUX, DALL·E, Stable Diffusion, and Leonardo, with goal modes when the still is a reference. Guests get about 3 image runs per day; a free account raises that to about 5. Soft start: https://promptmake.net/image. The steps below cover blank briefs and photo-first jobs. Keep lighting words stable across those rounds so you can see what the height change did.

From a blank brief to first paste

1. Write the job in one sentence: "45-degree duck plate, walnut," or "overhead ramen, dark bowl," or "cookbook lasagna cut, oak and linen."
2. Expand into the stack: dish, plating, camera height, light, surface. Add one short medium lock you will not edit this round (restaurant food photography or cookbook photography).
3. Pick canvas: --ar 4:5 for menu and print pages, --ar 1:1 for overheads, --ar 16:9 for a site hero. FLUX and SDXL get the matching preset or pixels.
4. Format: Midjourney phrases plus --style raw --v 7. FLUX full sentences. SDXL tokens plus a short negative against plastic food. GPT Image a short paragraph.
5. Generate once. Check height and table first. If the burger is a map, you used overhead on a height dish. If the bowl looks like a product packshot, you left the table blank. Change one fact: height, or surface, or key. Leave the dish nouns alone until the venue reads.

Save the winning prompt beside the PNG with the kit in the filename (menu_duck_45_walnut_v3.png) so the next teammate does not reopen an overhead and wonder why the bun vanished.

From a plate photo (Recreate, Variation, Change Style)

Upload a still that shows the plate you need: a pass photo from service, a cookbook test shot, a prior AI still with the right table. On PromptMake /image, pick Midjourney, FLUX, DALL·E, Stable Diffusion, or Leonardo as the paste target. Recreate Exactly when you want near-match language and will confirm height and surface. Create Variation when you need sibling dishes inside the same venue. Change Style when the plate stays but the medium moves (menu still to cookbook linen). Adjust Lighting when framing is fine and the key is wrong; that mode is for lamps. After copy-out, read the plate clause. Insert "45-degree" or "overhead" or "dark walnut" if vision wrote "food photography" with no height. Split a height rewrite and a lighting rewrite into separate runs; one overloaded pass muddies the draft. Pair the output with the food photography use case at https://promptmake.net/use-case/food-photography if you want tip language beside the generator.

Common mistakes with food photography prompts

Delicious food with no height or table. The model guesses. Write dish, plating, camera height, light, and surface.

Two heights in one line. Overhead flat lay plus 45-degree burger hero plus macro glaze is three cameras. Pick one kit per generate.

Treating --ar as a viewpoint. Aspect ratio is the box. Overhead and 45-degree are camera facts. Set both.

Ceiling light on a texture dish. Top light flattens sauce and bun shadow. Name window side light or a single key from camera left.

White seamless on a restaurant brief. Packshot sweep belongs to product work. Name walnut, slate, or linen.

Garnish pile and taste adjectives. Two garnish nouns beat a paragraph of juicy, mouthwatering, gourmet. Steam is a clause. Shine is a light fact.

Chapter drift in a cookbook. Marble on page 12 and oak on page 40 reads as two books. Lock surface and Kelvin for the chapter.

Changing height and lighting in one rewrite. You cannot tell which edit worked. Lock the lamp while you hunt for the right camera.

When PromptMake /image helps food photography prompts

Blank-page prompts work when you know the kit and the stack. Photo-first workflows win when a still defines the plate: a pass photo, a test shot from a recipe day, a menu still you want to sibling. PromptMake /image runs vision with model-specific formatting and goal modes: Recreate Exactly, Change Style, Adjust Lighting, Create Variation. Same upload, Midjourney dialect one run, FLUX the next. Soft path: https://promptmake.net/image. After the draft lands, you own the plate line: insert height, surface, and Kelvin before you spend generator credits.

Use chat when you explore GPT Image wording and need a fast "move to 45 degrees" tweak in one thread. Use /image when a still exists and you need recreate or variation text aimed at Midjourney, FLUX, DALL·E, Stable Diffusion, or Leonardo. Keep this article on food photography prompts so your first grid matches the plate you meant.

FAQ

What are food photography prompts?

Food photography prompts are text that tells an image model how to shoot a plate: dish, plating, camera height, light, and table surface. You paste them into Midjourney v7, FLUX, Leonardo, SDXL, or GPT Image so the frame matches a restaurant menu still or a cookbook spread. Name angle and surface, because a prompt that stops at delicious food leaves the table to chance. PromptMake /image can recover that language from a plate photo, then you confirm garnish and steam before you generate.

How do I write food photography prompts for Midjourney v7?

Lead with the dish, then add plating, camera height, light, and surface in compact phrases. End with --ar for the canvas, --style raw for photographic food, and --v 7. Example core: 45-degree camera, window side light from camera left, dark walnut table --ar 4:5 --style raw --v 7. Style references such as --sref hold look, so you write the plate clause by hand.

Should I prompt overhead or 45-degree for restaurant plates?

Overhead flat lay fits bowls, pasta, and table spreads where the layout is the story. A 45-degree or eye-level camera fits burgers, cakes, stacked mains, and drinks with height you need to show. Pick one height per generate. Mix both in one line and the model splits the difference into a mushy tilt.

How do cookbook stills differ from restaurant menu shots?

Restaurant menu stills isolate one plated dish on a locked surface with negative space and a short garnish story. Cookbook spreads add repeating props and linen so the chapter feels like one shoot. Keep the same wood, ceramic, and Kelvin across a chapter. Change the dish nouns and leave the table and key.

Which model handles food texture in 2026?

FLUX family builds (name the variant your host shows) handle photoreal food texture well. Midjourney v7 adds editorial polish that cookbook art directors ask for. GPT Image follows a conversational rewrite if you need "move to 45 degrees" in chat. Ideogram v3 helps if a recipe title must sit in frame; you write the plate stack.

How do I add steam and garnish without fake shine?

Name steam rising or condensation on glass as a short clause. Name one garnish noun, basil or sesame or citrus zest. Leave taste-adjective piles out of the brief; those words push plastic shine. Add "styled food photography" or "cookbook photography" as a medium lock, then generate.

How do I start for free today?

Write one sentence with dish and height ("45-degree duck plate" or "overhead ramen bowl"), expand it with the five-part stack, and paste a Midjourney v7 or FLUX template from this guide. If you hold a plate photo, open https://promptmake.net/image, pick your model, run Recreate Exactly or Create Variation on the free quota (about 3 guest runs, about 5 registered), then insert missing surface and Kelvin before you generate. Save the winning prompt beside the PNG as the kit seed for the next dish. Extra tip language lives on https://promptmake.net/use-case/food-photography.